Validation of the selected metrics and the results

13 experts from different MSs were contacted with a proposal to validate the used theoretical framework and energy poverty metrics. It is important to note that this was prior to the testing phase, which provided key insights into the application of different metrics. The names of the experts and their relevance to the topic of energy poverty are listed in Annex 4, along with the approved minutes of the interviews.

Summary of comments to theoretical framework

In general, stakeholders agreed that the energy poverty definition presented²⁸ is suitable. An important comment made regarding the proposed definition was that “a common mistake in the literature across Europe is to say that energy poverty only refers to heating, when in fact it refers to all energy services in the home, including cooling; by saying all energy services in the home, there is no room for

confusion or misinterpretation.” (Harriet Thomson)

Regarding the different approaches to monitor and measure energy poverty, the expenditure-based approach was mentioned by all the interviewees as the most suitable to measure energy poverty at present, although it has certain disadvantages. The rationale set out for using different approaches was agreed.

Summary of comments to energy poverty metrics

Supporting indicators are important to measure energy poverty, even though they are not easily compared across Member States. The set of supporting indicators could be broader and could include the indicators related to energy market liberalisation, competition in energy market, frequency of tariff switching by users, social assistance aspects, and buildings’ energy efficiency. The supporting indicator help explain the phenomenon of energy poverty and are thus also correlated with the energy poverty metrics that are studied in this report. Chapter 3 performs a systematic statistical analysis of these correlations by means of econometric regressions.

The main strength of the expenditure based metric is that it accurately captures the extent of energy poverty, referring to “required” expenditure approach. As stated by interviewees, its main weakness is that it takes many data to produce a “required” expenditure metric. Overall, expenditure based metrics also have some further disadvantages:

 The approach is not standardized;

 The metrics are not comparable across Member States;

 They do not cover hidden energy poverty²⁹;

 They do not consider general living conditions³⁰ and evolution of energy prices; and

 They do not take into account actual energy needs of the households³¹.

Regarding expenditure thresholds for these metrics, the 10% approach does not appear to be objective and comparable across MSs. There is no preference regarding threshold, as income levels differ greatly across MSs, and change over time. As such, the threshold should probably not be fixed and might be different for different MSs or MS groups. There is a value in using relative thresholds (i.e. twice-median expenditure). Also using a threshold related to minimum income could be comparable across Member States.

Consensual based metrics have an advantage of easier implementation, as there is a standardized survey basis across MSs (SILC). It provides an insight of the energy poverty issues, based on information about actual energy needs of households. Its main weaknesses are that it is difficult to interpret because of its subjective nature, the survey is not detailed enough, the answers can underestimate

28 Energy poverty was defined as a situation in which individuals or households are not able to adequately heat or meet other required energy services in their homes at affordable cost.

29 This was an expert opinion, though in the context of this study hidden energy poverty is one of the expenditure-based metrics considered. Further, it is important to note that accounting for hidden energy poverty (and similar issues such as self-disconnections) is one of the main advantages of using required energy expenditure.

30 This was an expert opinion, though in the context of this study MIS indicators do take into account general living conditions.

31 This was an expert opinion, though in the context of this study expenditure-based metrics could take this into account if based on detailed building stock information.

energy poverty, because the households are not willing to admit they are in a difficult situation with paying for energy services, and it depends on the group of households chosen for the survey.

Outcome based approach is the least preferred as the primary metric of energy poverty. Its main weakness is the difficulty of implementation. It is difficult to identify clearly the energy poverty outcomes; further, health and social related outcomes are too complex to measure. Also, there is an uncertainty concerning capturing the actual state of the issue, as it is concentrated only on the outcomes, and does not consider the causes of energy poverty.

Hidden energy poverty could be included with the use of consensual based indicators, for example the temperature level of household. In places where temperatures are milder (relative to the national median), household energy expenditure is expected to be lower, but that should not be necessarily seen as a problem. A comprehensive indicator of hidden energy poverty is presented in the Belgian Energy Poverty Barometer. The households whose energy expenditures were too low were identified by taking into account energy expenses of similar households (household composition and housing size).

The relative threshold for hidden energy poverty is defined for each household as half of the energy expenses of similar households with the same composition and housing size.

General conclusions:

 Expenditure based approach is probably at the moment best suited (with some reservations) to measure energy poverty across MSs, but experts argue that the ideal indicator would be an estimated amount of required energy, which is usually not feasible.

 Consensual based approach can be used to measure energy poverty as well, if the quality of survey will be improved.

 It is important to include supporting indicators when measuring energy poverty. The most important ones are those related to housing stock energy efficiency and energy market.

 When using expenditure based metrics, the threshold should be set relative to the actual distribution in the MS.

 It is reasonable to apply the selected indicators not only to low-income households.

3 Application of the Energy Poverty Metrics

In the previous chapter, energy poverty metrics and supporting indicators were chosen based on the literature and their use across MSs. However, this preliminary, qualitative analysis is not enough to conclude which indicators are the most suited for a wide application in the EU.

In order to decide which indicators are most adequate, these metrics and indicators were tested in selected MSs across various years and in different income groups using household-level data. The testing phase allowed an assessment of whether the list of indicators can be supported at EU-28 level.

Moreover, econometric analyses of the relationships between the chosen energy poverty metrics and a group of supporting indicators allowed us to decide which ones are more strongly associated with the phenomenon of energy poverty in each of the MSs analysed.

The analysis was performed for four MSs with various energy poverty situations, climates and policy approaches: Spain, Italy, the Slovak Republic and the Netherlands. The selected four MSs take into account the differences in regulatory environments. Thus our choice includes one country with highly regulated end-user electricity tariffs (SK), one with completely market-based pricing for retail electricity (NL), and one with semi liberalized retail market (ES).

Different regulatory environments

In particular, the electricity market in the Netherlands has been fully open to competition since 2004, with four major players in the country. As such, the electricity market in this country can be seen as a liberalized retail market. Furthermore, the retail prices of electricity are not regulated in the Netherlands per se but suppliers are obligated to report all price changes. In this regard, the authorities have the power to reduce prices as suppliers cannot provide sufficient justification for the amounts charged.

A similar energy market is seen in Slovakia, where the wholesale activities were fully liberalized in 2005. As such, there are no price regulation at this level. Furthermore, in 2012, Slovakia adopted laws for the further liberalisation and harmonisation of the energy market in the country. Nevertheless, the largest power generating company (Slovenské elektrárne) had still a market share of almost 78 percent in 2011. Moreover, the retail prices in Slovakia are still regulated through “price caps” for all households and small industrial users. The regulatory cycle in Slovakia is 5 years, the elaboration phase of which includes a consultation process including all market participants which represents additional step to enhance the transparency and predictability of the regulatory framework.

On the contrary, Italy has a free market which aims for free electricity trading for all commercial clients since July 2004 and a complete opening of the market for private customers from July 2007. However, the standard offer market remains concentrated, despite the numerous active suppliers, with three main operators.

Lastly, Spain has a highly regulated end-user electricity tariff. The electricity market in Spain was integrated with the Portuguese electricity market in 2007. There is a relatively high degree of concentration and vertical integration in the Spanish electricity market as a few players have a dominant role.

Source: EC, country reports; European Energy Market Reform, country profiles

The analysis was two-fold:

1. Testing the viability of the set of indicators based on data availability and the possibility to calculate the selected indicators at MS level; and

2. Analysing the data and assessing whether the chosen set of indicators is optimal. On this phase, we assessed 1) how energy poverty metrics correlate with each other, and 2) how different measures are influenced by a number of supporting indicators.